Dell PowerEdge XE9680 - Dual Xeon Platinum 8558, 8 × H200 GPU, 2,6 PB NVME & NDR Networking
  • Catégories de produits:Serveurs
  • Numéro de pièce:Dell PowerEdge XE9680
  • Disponibilité:In Stock
  • État:Tout neuf
  • Caractéristiques du produit:Prêt à expédier
  • Commande minimale:1 unité
  • Le prix catalogue était:$349,999.00
  • Votre prix: $342,858.00 Vous économisez $7,141.00
  • Discuter maintenant Envoyer un e-mail

Respirez tranquillement. Retours acceptés.

Expédition : l'expédition internationale d'articles peut être soumise à des formalités douanières et à des frais supplémentaires. Voir les détails

Livraison : veuillez prévoir un délai supplémentaire si la livraison internationale est soumise à un traitement douanier. Voir les détails

Retours : 14 jours pour les retours. Le vendeur paie les frais de retour. Voir les détails

Livraison GRATUITE. Nous acceptons les commandes d'achat NET 30 jours. Obtenez une décision en quelques secondes, sans affecter votre crédit.

Si vous avez besoin d'une grande quantité de produits Dell PowerEdge XE9680, appelez-nous via notre numéro gratuit au Whatsapp: (+86) 151-0113-5020 ou demandez un devis via le chat en direct et notre responsable des ventes vous contactera sous peu.

Title

Dell PowerEdge XE9680 – Dual Xeon Platinum 8558, 8× H200 GPUs, 2.6 PB NVMe & NDR Networking

Keywords

XE9680, Xeon Platinum 8558, HGX H200, NVMe SSD, ConnectX-7, 25GbE SFP+, 2U Rack Server, 2800W Titanium PSU

Description

The Dell PowerEdge XE9680 is a flagship, GPU-dense server platform engineered for large scale AI training, HPC, and inference workloads.

It supports dual Intel Xeon Platinum CPUs and up to **8 NVIDIA HGX H200 GPUs** (141 GB each), connected via NVLink and designed for extreme computational throughput.

Your build includes **2× Intel Xeon Platinum 8558** processors, which deliver very high core count and power envelope, making this system capable of heavy parallel compute tasks and massive model training.

Combined with **32 × 64 GB DDR5 ECC RDIMM** (for 2,048 GB RAM), the memory capacity provides broad room for large models and data-intensive operations.

In terms of storage, the system includes a BOSS-N controller with 2 × 960 GB M.2 in RAID-1 for boot / system, plus **2 × 3.84 TB U.2 NVMe SSDs** for high-speed storage.

This combination of PCIe Gen4 / NVMe devices ensures extremely fast I/O latency and throughput for both system tasks and data handling. (Your spec plus XE9680 supports many more front NVMe/SAS/SATA slots for expansion up to large capacities.)

Networking is outstanding: **8 × ConnectX-7 Single-Port NDR VPI OSFP PCIe FH** cards plus a **10-GbE Base-T NIC (X710-T4L)** and 1 GbE LOM provide both ultra-high bandwidth (for GPU-to-GPU data transfer, cluster interconnect) and flexible downstream network connectivity. 

Power delivery is robust: the spec calls for multiple **2800W Titanium Power Supplies** in redundancy (3+3 FTR or similar scheme) to support both the high power demands of the GPUs and all other components under full load.

The airflow & cooling is configured accordingly. You get “rack-mount 2U sliding rails” included, and the server comes from factory with no OS — ideal if you want to install your own hypervisor or Linux distro.

Key Features

  • 8 × NVIDIA HGX H200 141 GB GPUs (SXM) for massive GPU memory bandwidth and NVLink interconnect.
  • Dual Intel Xeon Platinum 8558 CPUs for high core/thread count compute.
  • 2 TB+ DDR5 ECC memory (32×64 GB) — large capacity for AI/ML HPC workloads.
  • Boot / OS via BOSS-N + 2 × 960 GB M.2 SSD in RAID-1, plus 2 × 3.84 TB NVMe U.2 SSDs for high-speed storage.
  • High-performance networking: 8 × ConnectX-7 NDR, plus 25 GbE SFP+ and 1 GbE options.
  • Redundant high-wattage Titanium PSUs (e.g. 2800W, 3+3 or 6 unit redundant) to ensure reliability under full GPU load.
  • No OS preinstalled — flexibility for custom deployment.
  • Remote management via iDRAC9, supports firmware / system monitoring.
  • 6U (or size matching spec) rack profile from Dell XE9680 family, with sliding rails included for rack mounting.

Configuration

ComponentSpecification
ModelDell PowerEdge XE9680
Processors2 × Intel Xeon Platinum 8558
GPUs1 × NVIDIA HGX H200 assembly (8 × SXM 700W GPUs, 141 GB each)
Memory32 × 64 GB DDR5 ECC RDIMM (2,048 GB)
Boot / OS StorageBOSS-N controller + 2 × 960 GB M.2 SSD (RAID-1)
Data Storage2 × 3.84 TB U.2 NVMe SSDs
Network Cards8 × ConnectX-7 NDR VPI OSFP + 1 × X710-T4L Quad-Port 25 GbE Base-T + Broadcom 5720 Dual-Port 1 GbE LOM
Power Supplies6 units 2800W Titanium (3+3 or 6 unit redundant layout)
Mounting / ChassisRack 2U (or size/6U profile depending XE9680 frame), sliding rails included
OSNo OS preinstalled

Compatibility

The XE9680 supports 8 × HGX H200 GPUs SXM5, fully NVLink interconnected, as per Dell’s spec sheet. 

Memory: Up to 4 TB using DDR5 RDIMMs across 32 slots. Your 2TB config is within this. Supports up to ~5600 MT/s on 5th Gen Xeon or lower speeds depending on CPU generation. 

Storage fronts: The XE9680 supports front NVMe/SAS/SATA slots, U.2/E3.S drive expansion up to about 122.88 TB depending on drive sizes.

Networking: The available PCIe Gen5 slots plus support for ConnectX-7 and Base-T NICs ensure full bandwidth for GPU-to-GPU and network traffic. Make sure firmware and NIC drivers match your networking stack.

Usage Scenarios

Large Model Training — With 8 × H200 GPUs, massive memory, and NVLink, excellent for deep learning training (LLMs, multi-modal models) where GPU memory is key.

Inference at Scale — High-bandwidth network (ConnectX-7 NDR) plus fast NVMe storage makes it suitable for deployment of inference clusters with low latency.

HPC Workloads & Simulations — Compute, memory, and GPU interconnect capabilities support simulation, modeling, and scientific compute tasks.

Data-Intensive Storage Tasks — NVMe SSDs plus large memory allow handling of datasets, analytics pipelines, caching layers, etc.

Frequently Asked Questions

Q1: How many PCIe Gen5 slots are left for expansion after GPU modules?
A: The XE9680 typically has up to 10 PCIe Gen5 x16 slots (varies depending on GPU / riser configuration). Some slots are used by GPUs; the rest are available for NICs, NVMe, etc. 


Q2: Are 8 × HGX H200 GPUs supported with Platinum 2800W PSUs?
A: Yes — Dell’s spec for XE9680 supports 2800W Titanium or equivalent redundant PSU configurations to power 8 × HGX H200 GPUs plus other components. 


Q3: Can I run Linux / Ubuntu / VMware on this build?
A: Yes — XE9680 supports major Linux distros (Ubuntu LTS, RHEL, SUSE) and VMware ESXi. Drivers for GPU, NICs and storage controller must be correctly matched.


Q4: What cooling and power redundancy considerations are required?
A: Ensure full airflow, high ambient cooling. Use redundant PSU layout (e.g. 3+3 or 6 unit redundancy). Make sure facility power infrastructure supports the load, especially with 8 × 700W GPUs plus CPUs and NVMe SSDs. 

PRODUITS LIÉS À CET ARTICLE
Fusionserver 5288V5 - Dual Xeon Platinum 8570, 1 To DDR5, stockage NVME / SATA mixte, Dual 10 GBE Recommander
Inspur NF5280A7-G7 - Dual Epyc 9654, 512 Go DDR5, NVME mixte et stockage SATA, DAL 1,3 KW PSU Recommander
Lenovo ThinkSystem SR250 V3 - Xeon E-2434, 32 Go de RAM, Dual 960-Go SSDS SSDS, RAID 9350-8I Recommander
Dell 7626 Rack Server - EPYC 9354 32-Core, 128 Go DDR5, 4 × 7,68 TB NVME, Dual 25 GBE NICS Recommander
HPE ProLiant DL380 Gen10 - Dual Xeon Gold 6148, 256 Go DDR4, 24-SFF, P816I-A RAID Recommander
Huawei Fusionserver 2288H V7 - Dual Xeon Gold 6548N, 1TB DDR5, 184TB NVME / SSD, 1500W PUISSANCE ENVANT Recommander
HPE ProLiant DL380 Gen11 Rack Server - Xeon Gold5416s, 128 Go DDR5, 6 × 3,84TB SSD, MR408i - O Recommander
Lenovo ThinkSystem SR650 - Double Intel Xeon Gold6230 Dual-CPU Server avec 256 Go de RAM et 8 × 1,2 To de stockage SAS 10K Recommander